Temperament and Stress

Over the years, we have been investigating differences in infants’ and children’s response to stress and the relation of stress differences to different aspects of their development. Of particular interest to us has been cortisol measures of stress which now can be readily obtained from small amounts of saliva. Our work has examined age changes in infants’ cortisol and behavioral responses to stress, as well as the role of environmental factors in these changes. In addition, our work has assessed the relation of cortisol response to emotional expression; in particular, the self-conscious emotions that emerge in the second and third year of life. Finally, given the long-established link between stress and adverse health outcome, our work has addressed whether a high cortisol response to stress is a marker for less optimal patterns of development in children born with different neonatal environmental and medical risk conditions.

Stress and Health in Children

The link between mind and health has long been recognized (Mens sama in corpore sano). Research in behavioral medicine is replete with examples of the effects of stressful life events, lifestyles, and personality attributes on subsequent illness in adults. Our past work examined the relation of newborn and young infants’ behavioral response to stress to various health outcomes. A high response to stress in newborns was related to fewer health problems at older ages, whereas a high response to stress by 2 months was related to more heath problems at older ages. Consistent with our cortisol results, we believe that a high stress response in newborn and young infants is a sign of more optimal functioning (i.e., more vigor), whereas a high stress response in older infants is a sign of less adaptive functioning.

Stress and Coping

We have been investigating differences in infants’ cortisol and behavioral responses to stress. From birth there are large individual differences in children’s response to stress. Using inoculation as the stressor, we have examined age change in infants’ cortisol response to stress. There is an age-related decline in magnitude of cortisol response between 2 and 6 months, but no further age change in magnitude of cortisol response between 6 and 18 months of age. Stable individual differences in cortisol response are apparent only from 6 months, remaining stable through 18 months of age. Whereas a high cortisol response indicates more optimal functioning in the newborn period, a high cortisol response indicates less optimal functioning by 6 months of age. Together these findings indicate considerable developmental change in adrenocortical functioning during the first 6 months of life, with the more mature organization of the adrenocortical system not present until approximately 6 months of age. Comparable results were obtained when we compared infants whose inoculations involved one versus two injections. In comparing infant cortisol and behavioral stress responses, we have found, at best, only modest relations between the two types of response. Since individual differences in the mode of expression of stress are likely, from a purely measurement point of view, both cortisol and behavioral measures are needed to adequately assess differences in stress. The absence of behavioral indications of stress does not necessarily mean that infants are not stressed.

Our work also has addressed environmental and constitutional variables associated with differences in infant cortisol response to stress. We explored the role of socialization and genetic factors in infant response to stress by contrasting the stress responses to inoculation of Japanese and Caucasian American infants. Observations of the infants’ behavioral distress indicate that the Caucasian infants are more reactive than the Japanese infants. On the other hand, Japanese infants show a greater cortisol response, suggesting higher internal levels of stress. To also address whether socialization factors impact infant responses to stress, we assessed the relation of mothers’ soothing behavior to their infants’ stress responses. Maternal soothing to inoculation and to other more everyday episodes of infant distress (e.g., diaper change, dressing) was observed. There was no evidence that maternal behavior was efficacious in lowering infant cortisol or behavioral response to stress. Our results to date suggest that temperament might play a strong role in individual differences in infant stress response. In this regard, we have identified relations between infant temperament and cortisol response. For example, greater negative emotionality (i.e., a more difficult temperament) is associated with a higher cortisol response across 2 and 6 months of age.

Response to stress involves both a reactivity and a regulation component. Reactivity refers to the peak response following a stressor, whereas regulation refers to the subsequent dampening of response following the peak. In our recent work, we have begun to examine both reactivity and regulation differences in infants’ cortisol and behavioral responses to stress. Differences in cortisol reactivity and regulation were observed by assessing cortisol levels at regular intervals following the stressor; that is, during the period of time that peak response and response dampening would be expected to occur. Results show that reactivity and regulation are unrelated for both cortisol and behavior. The independence of reactivity and regulation suggests that measures of both are needed to more completely characterize infant cortisol or behavioral response to stress. Moreover, there is considerable variation in the timing of the peak cortisol response, suggesting that obtaining only a single post-stressor cortisol sample might not provide a sensitive measure of cortisol reactivity in individual cases.

Stress and Self-Conscious Emotions

We have been investigating the relation of children’s cortisol response to their expression of the self-conscious emotions of pride, shame, and embarrassment following task success and failure. In a sample of 4-year-old children, we found that a high cortisol response is related to children’s expression of shame and embarrassment following failure, but that cortisol response is unrelated to children’s expression of pride following success. Longitudinal data suggest that a high cortisol response from early infancy (a relatively higher cortisol response by 6 as opposed to 2 months of age) is related to shame and embarrassment, but unrelated to pride. We believe that a high response to stress makes it more likely that children will engage in self-focused attention. Self-focused attention refers to attention directed internally toward one’s feelings and thoughts as opposed to externally toward the environment. Given that negative self-evaluation is stressful, the intense and prolonged self-focused attention following failure brought on by a high response to stress increases the likelihood that children will attribute the negative outcomes internally to the self rather than externally to the situation and, therefore, that they will express shame and/or embarrassment. Consistent with this view, in previous work we found that a high cortisol response from infancy is related to the earlier emergence of visual self-recognition. Self-recognition reflects the capacity for a self meta-representation that makes self-focused attention possible.

Stress and Prenatal Drug Exposure

There is reason to believe that prenatal exposure to drugs might adversely impact children’s cortisol response to stress. In one study, we examined the effect of prenatal alcohol and cigarette exposure on infants’ cortisol response to inoculation at 2 and 6 months of age. Cortisol response at 2 months was lower for the non-exposed than exposed infants, whereas cortisol response at 6 months did not differ between the exposed and unexposed infants. The 2-month group difference in cortisol response reflected a higher pre-stressor cortisol level in the exposed infants. In a second study, we examined the effect of prenatal cocaine exposure on infants’ cortisol response to inoculation at 2, 6, and 12 months of age. Across age, we found a sex difference in the effects of prenatal cocaine exposure on cortisol response. A high cortisol response to stress associated with prenatal cocaine exposure is present in boys, but not in girls. This finding is consistent with recent results from other prenatal cocaine exposure work at the Institute that boys might be more adversely affected by prenatal cocaine exposure than girls.

Stress in HIV-Infected Children

Prenatal exposure to HIV or to antiretroviral therapeutic agents such as AZT might affect the integrity of the hypothalamic-pituitary-adrenal axis. We are studying reactions to the stress of inoculation in infants exposed to HIV and AZT in utero. Our results indicate a high cortisol response to stress in exposed infants which, relative to controls, already is present by 2 months and is maintained through 12 months of age. It is hoped that this research will facilitate our understanding and management of the disease process in infected children, as well as our knowledge of the impact of these exposures on increasing numbers of uninfected children.